HKChat racks up 90,000 users in week one, tops Hong Kong App Store

HKChat, Hong Kong's AI assistant, hit 90,000 users in week one and took No. 1 on the local App Store. It helps with bus times, weather, and laws in three languages.

HKChat's fast start: 90,000 users in week one and the top spot on Hong Kong's App Store

Hong Kong's first locally built AI chatbot, HKChat, is off to a strong start. The app added around 90,000 users in its first week and reached the No. 1 spot among free apps on Apple's Hong Kong App Store, based on data from Sensor Tower.

The pitch is simple: answers for daily, local needs. HKChat surfaces real-time bus arrivals, weather updates, and information on Hong Kong laws and regulations. It supports Cantonese, Mandarin, and English from day one.

Built by the Hong Kong Generative AI Research and Development Centre (HKGAI), the app runs on HKGAI V1, a large language model developed using China's open-source DeepSeek models. The project ties into government ambitions to strengthen the city's tech footprint.

"Our goal is to provide Hong Kong residents with a good AI assistant," said Guo Yike, HKGAI director and provost at the Hong Kong University of Science and Technology (HKUST). "I think this might be the first such initiative anywhere in the world."

Why this matters for product and engineering teams

Local-first utility wins attention. HKChat prioritizes high-frequency city tasks (transport, weather, regulations) over generic chat-clear value beats broad feature lists.
Multilingual support is a core feature, not an add-on. Building for Cantonese, Mandarin, and English expands reach and reduces friction across user groups.
Open-source foundations can speed delivery. Using DeepSeek as a base model suggests faster iteration and more control over deployment and cost profiles.
Distribution signals are strong. Topping the App Store indicates demand for practical AI assistants that plug into real-time public data.

Early traction by the numbers

~90,000 users in the first week after a low-key release.
More than 8,000 new users per day across the last two days reported.
No. 1 free app on Apple's Hong Kong App Store.

What to watch next

Quality of real-time answers: accuracy on bus timings, consistent weather feeds, and up-to-date legal information.
Latency and reliability at scale: traffic spikes will stress data pipelines and model response speeds.
Privacy, data residency, and compliance: especially for government-adjacent data and multilingual inputs.
Integration ecosystem: whether HKChat exposes APIs or partners with transport, utilities, and city services.

Practical takeaways for teams building local assistants

Start with 3-5 high-volume use cases and wire them to reliable tools (transit, scheduling, official docs).
Choose an open model baseline that meets your latency, cost, and compliance needs; add retrieval and tool use early.
Instrument feedback loops from day one (thumbs up/down, "was this useful?") and ship weekly improvements.
Treat multilingual support as a product requirement if your market demands it-design prompts, evaluation, and QA accordingly.